ONTO
Onto Innovation
Summary
What they do:
Build specialty inspection and metrology systems — the Dragonfly platform — that detect defects and measure critical dimensions on advanced packaging substrates, HBM stacks, and 2.5D/3D interconnects. They sit between the fab and the packager, catching the yield-killing defects that emerge when you start stacking dies and bonding chiplets.
Why they matter:
Advanced packaging is the new bottleneck in AI chip production. Every CoWoS interposer, every HBM stack, every chiplet-based GPU needs inspection at the packaging level — and Onto is the niche specialist displacing incumbent tools at major HBM manufacturers. The Dragonfly G5 just won double-digit system orders from a leading HBM maker, beating the established tool of record on both resolution and throughput.
Recent performance:
Full year 2025 revenue $1.005B (first time crossing $1B), up 1.8% YoY. Q4 2025 revenue $267M, non-GAAP gross margin 54.6%, EPS $1.26. Q1 2026 preannounced at $292M, beating the $275–285M guide and $280M Street consensus. Q2 2026 guided $320–330M vs. $303M consensus. Stock ~$291, market cap ~$14.5B. Stifel upgraded to Buy with a $350 target.
Our Verdict
Niche advanced packaging inspection specialist riding the CoWoS/HBM wave — Dragonfly G5 displacing incumbents, $240M+ HBM purchase agreement locked in, 30%+ packaging growth in 2026. Risk is KLA counter-punching.
Structural trends
Structural
73
/ 100
Moat
6/10
Specialty advanced packaging inspection niche, growing CoWoS/HBM relevance
AI Exp.AI Exposure
Embedded~25% AI
Play Type
EmergingAI Growth
~25%
Rel. Value
45
FAIRPriceLIVE
$261.59
+3.55%
Live via Yahoo Finance · refreshes every 5 min
Market Cap
$13.0B
P/E Ratio
94.4
P/S Ratio
12.9x
52W High
$263.39
52W Low
$85.88
52W Chg
204.6%
Beta
1.47
Onto Innovation makes the inspection and metrology tools that find defects in advanced semiconductor packaging. When TSMC builds a CoWoS interposer with twelve HBM stacks sitting next to an NVIDIA GPU die, or when SK Hynix bonds hundreds of layers in an HBM4 stack, someone has to verify that every bump, every via, every micro-interconnect is where it should be and free of defects. That someone is increasingly Onto.
The company was formed in 2019 from the merger of Nanometrics and Rudolph Technologies — combining front-end optical metrology with back-end packaging inspection. The flagship product line is Dragonfly, a 2D inspection and 3D metrology platform purpose-built for advanced packaging. The newest generation, Dragonfly G5, detects defects as small as 150nm and is designed to scale across multiple technology nodes. It just won a competitive evaluation at a major HBM manufacturer, displacing the previous tool of record and securing double-digit system orders plus a volume purchase agreement worth over $240 million through 2027.
Onto operates across two segments: specialty devices & advanced packaging (slightly over $500M in 2025 revenue) and advanced nodes (which more than doubled in 2025). The company has no debt, holds roughly $640M in cash, and runs at ~49 million shares outstanding. At $291 per share, the market cap is approximately $14.5B — small by semiconductor equipment standards but growing fast as advanced packaging inspection becomes a critical yield lever.
The customer base reads like a who's who of the AI chip supply chain: TSMC, Samsung, Intel, and SK Hynix all use Onto tools for critical metrology and inspection. The company is qualified on gate-all-around nanosheet architectures at all three leading foundries, giving it front-end relevance alongside its packaging dominance. But the growth story is packaging — advanced packaging revenue is expected to grow more than 30% in 2026, driven by HBM4 yield requirements and 2.5D AI packaging proliferation.
Supply Chain Dependencies
The Catch
Onto is a $1B revenue company trading at $14.5B market cap in a segment where KLA generates $11B+ in revenue and dominates process control. KLA has not attacked packaging inspection as a priority — yet. If HBM and CoWoS packaging inspection becomes a $2–3B addressable market (which is exactly the thesis for owning Onto), it becomes worth KLA's attention. KLA has deeper customer relationships, larger R&D budgets, and a track record of crushing niche players when it enters a market seriously. Onto's competitive moat is real today — the Dragonfly G5 won a head-to-head evaluation — but it's a niche moat, not a structural one. The $240M purchase agreement locks in one customer, but it doesn't lock out KLA from the next customer or the next technology node.
If They Win
If the Dragonfly G5 becomes the standard inspection platform for HBM4 and next-generation 2.5D packaging at multiple manufacturers, Onto transforms from a $1B niche player into a $2–3B advanced packaging infrastructure company within three years. Every HBM stack, every CoWoS interposer, every chiplet-based AI processor would flow through Onto's inspection tools before it ships. The company would own the yield gate for the most critical and fastest-growing segment of semiconductor manufacturing. Revenue would compound at 25–30% through 2028, margins would expand as the tool mix shifts to higher-ASP G5 systems, and the stock would re-rate from "niche specialty" to "critical AI infrastructure." At that point, Onto either becomes an acquisition target for KLA, AMAT, or a larger equipment company — or it becomes the next KLA-in-miniature, dominating a segment that didn't exist at scale five years ago.
Others in Build the Machines That Make the Chip
Not financial advice. All scores generated via AI algorithms using public data.